Duality, Geometry, and Support Vector Regression

@inproceedings{Bi2001DualityGA,
  title={Duality, Geometry, and Support Vector Regression},
  author={J. Bi and K. P. Bennett},
  booktitle={NIPS},
  year={2001}
}
We develop an intuitive geometric framework for support vector regression (SVR). By examining when ǫ-tubes exist, we show that SVR can be regarded as a classification problem in the dual space. Hard and soft ǫ-tubes are constructed by separating the convex or reduced convex hulls respectively of the training data with the response variable shifted up and down by ǫ. A novel SVR model is proposed based on choosing the max-margin plane between the two shifted datasets. Maximizing the margin… CONTINUE READING

From This Paper

Figures, tables, and topics from this paper.
8 Citations
7 References
Similar Papers

References

Publications referenced by this paper.
Showing 1-7 of 7 references

Nonlinear Programming

  • O. Mangasarian
  • SIAM, Philadelphia,
  • 1994
Highly Influential
3 Excerpts

Crisp and C . Burges . A geometric interpretation of νSVM classifiers

  • D.
  • Proc . of Seventeenth Intl . Conf . on Machine…
  • 2000

Similar Papers

Loading similar papers…